When blood particles travel through an intravascular ultrasound imaging plane, the received echo signals decorrelate at a rate that is related to the flow velocity. In this paper, the feasibility of extracting blood velocity from the decorrelation function of radio frequency signals was investigated through theoretical analysis and computer simulation. A computer model based on the impulse response method was developed to generate the ultrasound field of a 30-MHz intravascular transducer. The decorrelation due to the scatterer displacement as well as other nonmotion related decorrelation sources were studied. The computer simulations show that the decorrelation function is linearly related to the lateral displacement. The monotonic relationship between correlation and displacement provides possibilities to estimate flow velocity with decorrelation measurements. Because of the complexity of the beam profile in the near field, assessment of local velocities requires detailed knowledge of the decorrelation at each axial beam position. Sources of signal decorrelation other than the lateral displacement may cause a bias in the decorrelation based velocity measurements. For localized decorrelation estimation, measurement variations in small range windows present a major challenge. An approach based on multiple decorrelation measurements should be adopted in order to reduce the variations. In conclusion, results of this study suggest that it is feasible to measure flow velocity by quantifying the decorrelation of intravascular ultrasound signals from blood.